1. Field of the Invention
The present invention is related to a defect inspection apparatus and a defect inspection method. More specifically, the present invention is directed to a defect inspection apparatus and a defect inspection method, for inspecting whether or not a manufactured product has been manufactured with a desirable shape and without any defect within a predicted range of a design work.
2. Description of the Related Art
As technical ideas capable of comparing/judging whether or not images obtained by photographing subjects to be inspected have desirable shapes in order to inspect defects, such techniques called as “template matching techniques” have been widely utilized. Then, generally speaking, defect inspections are carried out in accordance with the following sequential operations: That is, in such a case that a position where a defect may readily occur has been previously grasped due to reasons on manufacturing aspects and the like, an inspection position is found out by utilizing a so-called “template matching technique”, and thereafter, an inspection whether or not the defect is actually present at the position where the defect may easily occur is carried out in a precise manner.
As the conventional techniques related to the above-described template matching operations, such a technical idea is known which has been described in, for instance, a publication entitled “Basic and Application Ideas of Digital Image Processing Operation based upon Visual C#. NET & Visual Basic. NET-from Basic Concept up to Face Image Recognition-” written by Koich SAKAI, published by CQ publisher, first edition published on Sep. 1, 2003” and so on. The basic method as to the template matching technique has been described in this publication entitled “Basic and Application Ideas of Digital Image Processing Operation based upon Visual C#. NET & Visual Basic. NET-from Basic Concept up to Face Image Recognition” written by Koich SAKAI, published by CQ publisher, first edition published on Sep. 1, 2003”, in the third column of the fifth section, on pages 84 to 86. FIG. 5 and FIG. 6 of the above-described publication have exemplified such a correlative calculation result that (a) a face image of a girl is employed as an original image; (b) a left eye portion is cut out as a template; and while such an image formed by converting concentration of the original image is employed as an input image, a correlative calculation with respect to the template is carried out at each portion of the input image. Then, this publication has exemplified that (c) when a threshold value of a correlation coefficient is selected to be larger than, or equal to 0.99, only the position of the left eye cut out as the template is matched; and, however, when a threshold value of a correlation function is selected to be 0.8, a certain number of matching candidate positions may appear not only in the left eye, but also in the right eye.
Also, technical ideas capable of reducing calculation times of matching operations have been described in JP-A-63-211474, JP-A-2001-351102, and the publication entitled “Basic and Application Ideas of Digital Image Processing Operation based upon Visual C#. NET & Visual Basic. NET-from Basic Concept up to Face Image Recognition” written by Koich SAKAI, published by CQ publisher, first edition published on Sep. 1, 2003”, respectively. The matching calculation time reducing technique described in JP-A-63-211474 is given as follows: That is, the dimension of the image of the template, the sampling rate, and the candidate point area are decreased in the hierarchical manner. As a result, the matching calculation can be carried out in a higher speed, as compared with such a case that the template having the large size which has not yet been processed in the hierarchical manner is held. Then, another matching calculation reducing technique has been described in the publication entitled “Basic and Application Ideas of Digital Image Processing Operation based upon Visual C#. NET & Visual Basic. NET-from Basic Concept up to Face Image Recognition” written by Koich SAKAI, published by CQ publisher, first edition published on Sep. 1, 2003.” That is, in such a case that the binary image is employed, when the subtraction based upon the averaged value and the normalization can be omitted, “distance” is employed as the evaluation formula so as to shorten the calculation time, so that the process operation can be carried out in the high speed. Also, JP-A-2001-351102 has described such a method for performing the matching calculation of the binary image in the high speed. That is, in the case that the subject to be recognized is a relatively simple mode, while the circle is employed as the template, in which the width of the externally contacted rectangle for constituting the subject to be recognized is defined as the diameter, the matching calculation as to the binary image can be performed in the high speed.
Furthermore, JP-A-8-181053 has disclosed such a method capable of suppressing an erroneous recognition for positioning operation, which is suitable for an exposing apparatus which forms a large-sized liquid crystal display element. In this suppressing technique described in JP-A-8-181053, such a portion that a large number of errors may occur is masked which has been previously grasped within a template image; a candidate position is masked which also corresponds to a photographed image (input image); a calculation is carried out based upon a predetermined evaluation formula only as to such an area which is not masked, so that the position detecting precision can be improved.